Akkermansia muciniphila supplementation alters inflammatory profiles across diverse models of colitis

Abstract

Inflammatory bowel disease (IBD) is characterized by chronic intestinal inflammation and is thought to result from interactions among the immune system, environmental factors, and the gut microbiota in genetically susceptible individuals. Akkermansia muciniphila, a commensal bacterium has been reported to be depleted in individuals with IBD, although its precise role in intestinal inflammation remains unclear. This study examined the effects of A. muciniphila across multiple models of colitis, including dextran sulphate sodium (DSS)-induced colitis, the Mucin-2 knockout (Muc2^−/−) model of spontaneous colitis, and Trichuris muris-mediated infectious colitis. In a DSS recovery model, treatment with pasteurized A. muciniphila reduced the severity of inflammation. However, when administered prior to DSS exposure, both live and pasteurized bacteria did not significantly reduce inflammatory markers, suggesting limited preventive effects. In T. muris–infected mice, supplementation with live A. muciniphila increased Th2 and anti-inflammatory cytokine responses, reduced parasite burden, and enhanced gene expression of the mucin Muc5ac. Additionally, both live and pasteurized A. muciniphila alleviated spontaneous colitis severity in Muc2^−/− mice, indicating that these protective effects occur independently of Muc2. These findings expand understanding of the role of A. muciniphila in intestinal inflammation and highlight its potential as a therapeutic target for inflammatory intestinal disorders such as IBD.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1038/s41598-026-60105-9.

Introduction

Inflammatory bowel disease (IBD), characterized by chronic intestinal inflammation, is currently hypothesized to arise from the interplay between the gut microbiota, environmental factors, and the immune system in a genetically susceptible individual¹. IBD can be broken down into two symptomatically similar but physiologically different diseases, Crohn’s disease (CD) and ulcerative colitis (UC)². CD can manifest at any site in the gastrointestinal (GI) tract and is characterized by sporadic transmural inflammation². UC, on the other hand, is characterized by superficial mucosal ulcerations and varying degrees of continuous inflammation, which is largely confined to the colon³.

Though the initiation and pathogenesis of these debilitating diseases are currently unknown, evidence suggests that the resident microbial community of the gut may play a role in the inappropriate inflammatory response characteristic of IBD^1,3. Indeed, multiple studies have suggested that microbial diversity is diminished in IBD and that microbial shifts occur, predominantly characterized by increased levels of Proteobacteria and Actinobacteria and decreased abundance of Firmicutes and Bacteroidetes^3,4.

Within the GI tract, the gut barrier is the first line of defence, protecting the host from exogeneous threats, including microbial components. This barrier consists of epithelial cells and their junctional proteins, as well as the overlying mucus layer containing structural gel-forming mucins and antimicrobial peptides (AMPs)^5,6. Under pathophysiological stress, rearrangements and alterations in the various elements of this barrier enable the infiltration of unwanted luminal contents, notably microbes and microbial antigens, into the intestinal mucosa^7,8. Several studies have implicated impaired barrier function, including altered tight junction protein (TJP) expression and increased intestinal permeability^9–11, and, notably, alterations in mucin expression, glycosylation patterns¹², and mucus layer integrity⁶ in IBD. However, whether impairment of, or damage to, this layer is a primary causal factor or simply a consequence of exacerbated inflammation is still under investigation^9–11.

Akkermansia muciniphila, of the phylum Verrucomicrobia, is a commensal microbe that accounts for 1–5% of the human fecal microbiota^13–15. Evidence from a wide array of studies has supported the idea that this mucin degrader plays a role as a keystone species in the gut¹⁶ not only by directly supporting the health of the microbial community but also by interacting, either directly or indirectly, with the host and altering gut physiology. Despite its mucolytic nature, evidence suggests that this microbe and/or its component parts are particularly adept at fortifying and maintaining gut barrier integrity^15,17. These effects include, but are not limited to, increasing mucus thickness^15,18,19, increasing the number of mucin-producing goblet cells^18,20, stimulating mucus production²¹, as well as restoring appropriate tight junction expression and strengthening the epithelial barrier^22–24. A. muciniphila has also been shown to boost levels of anti-inflammatory T_reg cells and subsequent production of IL-10^25,26, and stimulate the production of AMPs such as regenerating islet-derived protein 3-gamma (RegIIIγ)¹⁵. Interestingly, several of these findings have been recapitulated, or even enhanced, with the use of pasteurized bacteria or isolated bacterial components²⁷.

Early research elucidating the role of A. muciniphila in inflammation was primarily focused within the context of obesity and metabolic syndrome^18,28,29. Increasingly, however, emerging evidence suggests this microbe and/or its component parts may have protective impacts in the context of colitis^{26,27,30–32}.

Though a clear-cut role and mechanism by which A. muciniphila may influence chronic inflammatory conditions have not been determined, there is evidence indicating that this microbe is depleted in patients with IBD and in chemical models of colitis^29,33–36, suggesting that A. muciniphila may have protective or anti-inflammatory effects which are lost in IBD³⁷.

Here, by utilizing three gut barrier-disrupting models of intestinal inflammation, including chemically-induced dextran sulphate sodium (DSS) colitis, the parasitic infection-induced model of colitis, Trichuris muris, and the spontaneous colitis developed in mucin-2 deficient (Muc2^−/−) mice, we explore the role of A. muciniphila in intestinal barrier function, host immune response, and the pathogenesis of intestinal inflammation across altered mucus environments. Our findings suggest this commensal bacterium, across various murine models of colitis, has the ability to modulate the host’s immunological profile and attenuate the severity of intestinal inflammation.

Results

Treatment with A. muciniphila after establishment of colitis reduces the severity of inflammation and enhances recovery in a DSS-induced colitis model

To explore the effects of A. muciniphila supplementation (AKK) in intestinal inflammation, its impacts on barrier function and to investigate whether the potential amelioratory effects of A. muciniphila on colitis were due to its metabolic properties or simply to the presence of the bacteria itself, live (Live AKK) or pasteurized (Past. AKK) bacterial samples or control brain-heart infusion (BHI) broth was administered by oral gavage to SPF C57BL/6 mice in both a recovery (Fig. 1a) and a prophylactic (Fig. 3a) experimental set up of the DSS colitis model.

Fig. 1.

Impact of the administration of live and pasteurized A. muciniphila in recovery from DSS-induced colitis. SPF C57BL/6 mice were administered low-dose (1.5%) DSS for five days. Starting on day 4 of DSS administration and continuing for the remainder of the experiment, mice received daily gavage of either live or pasteurized A. muciniphila, or control broth. Mice were allowed a recovery period of three days after the cessation of DSS. (A) shows the overall experimental layout. (B) shows macroscopic scoring analyzed at the time of sacrifice, and (C) shows disease activity index (DAI) which was recorded over the course of the DSS and recovery periods. (D) Histological scoring was also assessed. (E) shows H&E-stained representative micrographs and PAS-stained representative micrographs of colonic tissue. DAI is a compound score of weight loss, stool consistency, and fecal blood, with a maximum score of 12. Macroscopic score is a summation of colonic and rectal bleeding, stool consistency, and rectal prolapse at the time of sacrifice, with a maximum score of 10. In histological images, areas of damage and/or cellular infiltration are highlighted by a white arrow. Each value represents the mean ± SEM (n = 4–6 mice per group). *P < 0.05.

Fig. 3.

Effect of live and pasteurized A. muciniphila on markers of inflammation in a preventative supplementation model of DSS-induced colitis. SPF C57BL/6 mice received either live or pasteurized A. muciniphila, or control broth daily for 15 days by oral gavage. Following this, mice were administered 3.5% DSS for 5 days. (A) shows the overall experimental layout. Disease progress and severity of inflammation were assessed by (B) macroscopic scoring, (C) DAI, (D) histological scoring as well as levels of the pro-inflammatory cytokines (E), IL-1β and IL-6. (F) shows representative H&E-stained colon tissue micrographs in broth, live and pasteurized A. muciniphila-treated groups. DAI is a compound score of weight loss, stool consistency, and fecal blood, with a maximum score of 12. Macroscopic score is a summation of colonic and rectal bleeding, stool consistency, and rectal prolapse at the time of sacrifice, with a maximum score of 10. In histological images, areas of damage and/or cellular infiltration are highlighted by a white arrow. Data is represented as the mean ± SEM (n = 5–6 mice per group). *P < 0.05.

To elucidate the role of A. muciniphila in a recovery model of DSS-induced colitis, C57BL/6 mice were given low dose (1.5%) DSS for five days. Starting on day 4 of DSS administration, mice received either BHI broth or live or pasteurized A. muciniphila by oral gavage (Fig. 1a). This treatment continued until the end of the experiment. On day 5, DSS was removed, and mice were given access to sterile drinking water ad libitum for 3 days, which constituted the recovery phase. Disease activity index (DAI), a scoring system which includes stool consistency, body weight fluctuations, and blood in feces, over the course of the experiment, indicated similar levels of colitis were established until day 4 of DSS at which time a separation in the severity of colitis was seen between groups (Fig. 1c). Here, distinct differences were evident, with significance being reached on all 5 days of gavage treatment for the pasteurized A. muciniphila group in comparison to the broth group. It should be noted that before the AKK treatment period began, mice in the A. muciniphila groups, both live and pasteurized, showed increased weight loss during DSS administration (see Supplementary Fig. S1). To account for this observation, the DAI in Fig. 1 is reflective of stool consistency and blood in feces measures only (an additional analysis of DAI with body weight measures can be seen in Supplementary Fig. S1). Further analysis comparing within-group DAI scores of pre-gavage colitis (day 3) to the severity of colitis in recovery (day 7) was also performed (see Supplementary Fig. S2).

Upon sacrifice, macroscopic scoring (Fig. 1b), histological scoring (Fig. 1d,e) and the cytokines, IL-6 and TNFα (Fig. 2b,c) reflected the decreased severity of colitis observed throughout the experiment with regard to the pasteurized A. muciniphila group; significant decreases in all these measures were detected compared to controls. Mice receiving pasteurized bacteria also displayed significantly lower levels of IL-6 compared to those receiving live samples (Fig. 2b). Levels of TNFα were also significantly diminished in mice treated with live A. muciniphila compared to controls (Fig. 2c). No significant differences in IL-1β were detected between groups (Fig. 2d). Though no statistical significance was achieved when comparing levels of anti-inflammatory IL-10 (Fig. 2e) between groups, the pro- to anti-inflammatory ratio as represented by TNFα/IL-10 was markedly decreased in the pasteurized AKK-receiving groups compared with those receiving broth alone (Fig. 2f). In the broth group, higher histological scores were largely driven by infiltrating immune cells and disruption to the mucosal architecture.

Fig. 2.

Prophylactic impact of the administration of live and pasteurized A. muciniphila in acute DSS-induced colitis. SPF C57BL/6 mice were administered low-dose (1.5%) DSS for 5 days. Starting on day 4 of DSS administration and continuing for the remainder of the experiment, mice received daily gavage of either live or pasteurized A. muciniphila, or control broth. Mice were allowed a recovery period of 3 days after the cessation of DSS. (A) shows the number of PAS+ stained goblet cells per 10 crypts. Disease progress and severity of inflammation were assessed in colonic tissue by the pro-inflammatory cytokines, (B) IL-6, (C) TNFα, (D) IL-1β, and (E) the anti-inflammatory cytokine IL-10. The pro- to anti- inflammatory ratio as represented by TNFα/IL-10 is shown in (F). Markers of barrier function, including the tight junction proteins, (H) occludin and (I) zonula occludens-1, the antimicrobial peptides, (J) mouse β-defensin 1 and (K) RegIIIγ, and the major structural mucin of the gut, (L) Muc2, were assessed by qPCR. Data is represented as the mean ± SEM (n = 4–6 mice per group). *P < 0.05.

Evaluation of several markers of barrier function, including ZO-1, occludin, RegIIIγ and mouse β-defensin 1, were also assessed (Fig. 2h,l). No statistically significant differences were found between groups. Notably, no significant difference was observed with regard to the impact of either live or pasteurized A. muciniphila on expression of the main structural mucin of the gut, Muc2 (Fig. 2l). In addition, histological evidence (Figs. 1e and 2a) revealed no differences in the number of PAS+ stained mucin-producing goblet cells detected between groups.

To visualize the effects of live and pasteurized A. muciniphila supplementation on microbial composition and diversity, 16S rRNA sequencing of fecal samples was performed at the cessation of the recovery period (see Supplementary Fig. S3). At the genus level, beta diversity indicated distinct dissimilarities between microbial populations, in particular, between broth and pasteurized A. muciniphila-receiving groups (P = 0.015). Relative abundance measures suggest that although the absolute number of microbes may not be enhanced in the treated groups versus controls, the overall percentage of the microbial milieu occupied by A. muciniphila was enhanced.

Administration of either live or pasteurized A. muciniphila does not confer prophylactic effects in acute DSS colitis

In a preventative supplementation set up, A. muciniphila was gavaged for 15 days prior to inducing colitis via 3.5% DSS in autoclaved drinking water for five days (Fig. 3a). In this context, neither live nor pasteurized A. muciniphila administration proved beneficial. Markers, including macroscopic scoring (Fig. 3b), DAI (Fig. 3c), and levels of the cytokines IL-1β and IL-6 (Fig. 3e), proved similar between groups. In addition, histological evidence indicated that severe and consistent levels of colitis were established across all three groups (Fig. 3d,f). 16S rRNA analysis (Supplementary Fig. S4) suggested that in this prophylactic model, A. muciniphila was not able to establish a distinct foothold in the gut, post-DSS. These findings, in combination with the above work in a recovery setting, suggest that A. muciniphila’s amelioratory effects are highly context dependent.

Supplementation of live A. muciniphila promotes an environment favourable to T. muris expulsion in SPF C57BL/6 mice

To further explore A. muciniphila’s ability to influence intestinal inflammation within an altered mucin environment, we utilized the T. muris infection-induced model of colitis. The ability of T. muris-infected mice to successfully clear infection is heavily reliant on changes in the mucus environment, including altered mucus composition and goblet cell hyperplasia^38,39. SPF C57BL/6 mice were administered live (Live AKK) or pasteurized (Past. AKK) bacteria by oral gavage for 15 days starting one day before infection with ~ 300 T. muris eggs (Fig. 4a)⁴⁰. Mice were sacrificed 14 days post-T. muris infection.

Fig. 4.

Effect of A. muciniphila against T. muris infection-induced changes in intestinal epithelium and gut inflammation. Starting one day prior to the administration of ~ 300 T. muris eggs, SPF C57BL/6 mice received a daily gavage of either live or pasteurized A. muciniphila, or control broth for 15 days. Mice were sacrificed on day 14 post-T. muris infection. (A) represents a brief experimental outline. Infection severity and inflammation were analyzed by (B) worm burden as well as the cytokines (C) IL-4, (D) IL-13, (E) IL-10 and (F) IFN-γ. (G) shows the PAS-stained representative micrographs of colonic tissue from mice receiving broth, live A. muciniphila or pasteurized A. muciniphila and (H) shows the number of PAS+ stained goblet cells per 10 crypts. Genes involved in barrier function, including the mucins (I) Muc2 and (J) Muc5ac, the tight junction proteins (K) occludin and (L) zonula occludens-1, and the antimicrobial peptides (M) mouse β-defensin 1 and (N) RegIIIγ, were assessed by qPCR. Data is represented as the mean ± SEM (n = 4–5 mice per group). *P < 0.05.

Upon sacrifice, evaluation of infection severity and markers of the subsequent inflammatory consequences were performed. Here, mice administered live A. muciniphila displayed decreased worm burden (Fig. 4b), and minor up-regulation in the levels of IL-4 (Fig. 4c), and the anti-inflammatory cytokine, IL-10 (Fig. 4e), compared to controls. Measures of IL-13 (Fig. 4d) and IFNγ (Fig. 4f) did not show any notable difference between groups.

No differences in goblet cell number were observed between the groups in PAS+-stained colonic tissue; the live A. muciniphila-gavaged group displayed only minor signs of increased “fullness” (Fig. 4g,h). Parallel with these findings, Muc2 gene expression (Fig. 4i) did not prove significantly different between the groups, similar to findings in the DSS recovery experiment above. Of note, the live A. muciniphila treated group, however, when compared to both broth and pasteurized AKK treated groups, showed a significant upregulation in levels of Muc5ac (Fig. 4j), a gene which codes for a mucin not normally expressed in the colon but key in the latter stages of T. muris clearance^39–42. Across all other measured markers of barrier function (Fig. 4k–n), the group receiving the pasteurized microbe displayed no statistically significant differences between either the broth or live AKK-treated groups.

To measure the effects of live and pasteurized A. muciniphila supplementation on microbial composition and diversity, 16S rRNA sequencing of fecal samples was performed 14 days post-infection (see Supplementary Fig. S5). At the genus level, beta diversity indicated, in particular, significant dissimilarity between the broth and live A. muciniphila receiving groups (P = 0.015). Mice receiving live A. muciniphila also had distinctly enhanced (P = 0.003) alpha diversity compared to broth-receiving mice. In this model, relative abundance measures indicate live A. muciniphila was able to establish a foothold in the microbial milieu.

Live and pasteurized A. muciniphila ameliorates colitis progression and severity in spontaneous colitis developed in Muc2^−/− mice

Across both DSS and T. muris models, we observed a distinct lack of effect with regard to A. muciniphila’s impact on both goblet cell number and Muc2 gene expression. Thus, we set out to explore if the amelioratory effects observed in the aforementioned experiments are indeed independent of A. muciniphila’s known influence on mucin production and turnover^17,18,28,43 by utilizing the Muc2^−/− model of spontaneous colitis. These mice have severely diminished goblet cell numbers along with inhibited mucin production and develop spontaneous colitis as early as 5 weeks of age⁴⁴.

Live and pasteurized A. muciniphila were administered to Muc2^−/− mice via daily gavage for 15 days. Knockout (KO) mice and Muc2^+/+ receiving BHI broth, as well as Muc2^+/+ mice receiving live A. muciniphila, served as controls. It should be noted that at the start of the experiment, all Muc2^−/− mice had established colitis, the progress of which was monitored over the course of the experiment (Fig. 5a). Slight though not significant differences were observed in the measure of DAI between groups (Fig. 5b). However, mice gavaged with live or pasteurized A. muciniphila had reduced levels of both IL-6 and TNFα compared to their broth-receiving KO counterparts (Fig. 5c and d). Between the KO groups, IL-1β showed no remarkable differences (Fig. 5e). Here, though not statistically significant, TNFα to IL-10 ratios in both live AKK and pasteurized AKK-gavaged groups displayed a slight decline compared to those mice receiving the control broth, suggesting an overall “tamping down” of the inflammatory response may be in effect in A. muciniphila-treated groups (Fig. 5g). No difference in IL-1β, TNFα or IL-6 were detected between broth-receiving and AKK-receiving Muc2^+/+ groups, suggesting that A. muciniphila-mediated effects are colitis dependent.

Fig. 5.

Effect of live and pasteurized A. muciniphila on markers of inflammation in the Muc2^−/− model of spontaneous colitis. Muc2^−/− mice develop spontaneous colitis as early as five weeks of age. (A) Muc2^+/+ and Muc2^−/− (with established colitis) were administered a daily gavage of either live or pasteurized A. muciniphila, or control broth for 15 days. Over the course of the experiment, (B) DAI was recorded in Muc2^−/− mice to visualize disease progression over time. Upon sacrifice, the severity of inflammation was assessed by the cytokines (C) IL-6, (D) TNFα, (E) IL-1β and (F) IL-10. The pro- to anti- inflammatory ratio as represented by TNFα/IL-10 is shown in (G). DAI is a compound score of weight loss, stool consistency, and fecal blood, with a maximum score of 12. Data is represented as the mean ± SEM (n = 4–6 mice per group). *P < 0.05.

Markers of barrier function, occludin, ZO-1, mouse β-defensin 1, and RegIIIγ were analyzed in KO constituents (Fig. 6a–d). Of these markers, only occludin (Fig. 6a) was a statistically significant increase in the pasteurized A. muciniphila-receiving mice compared with controls.

Fig. 6.

Markers of barrier function and histological alterations in the Muc2^−/− model of spontaneous colitis in mice supplemented with live or pasteurized A. muciniphila. Muc2^−/− mice develop spontaneous colitis as early as five weeks of age. Muc2^−/− (with established colitis) were administered a daily gavage of either live or pasteurized A. muciniphila, or control broth for 15 days. Markers of barrier function, including the tight junction proteins, (A) occludin and (B) zonula occludens-1, and the antimicrobial peptides, (C) mouse β-defensin 1 and (D) RegIIIγ, were assessed by qPCR. (E) represents histological score, whereas (F) represents the number of PAS+ stained goblet cells per 10 crypts as shown in (H). (G) shows H&E-stained representative micrographs, (H) shows PAS-stained representative micrographs of colonic tissue and (I) shows Alcian blue-stained micrographs of colonic tissue. In histological images, areas of damage and/or cellular infiltration are highlighted by a white arrow. Data is represented as the mean ± SEM (n = 4–6 mice per group). *P < 0.05.

Between the groups of KO mice, H&E-stained colonic tissue micrographs revealed significant histological differences, particularly with regard to diminished inflammatory cell infiltration in both live and pasteurized groups compared to controls (Fig. 6e,g). Notably and quite surprisingly, we also observed distinct differences in both PAS+ stained goblet cell number and “fullness” in the live A. muciniphila receiving group compared with the broth-receiving group (Fig. 6f,h), suggesting A. muciniphila may be influencing the production of mucins other than MUC2 or glycosylation patterns within the gut. Intriguingly, no such cells were present when the same tissues were stained with Alcian blue to quantify goblet cells containing acidic mucins (Fig. 6i).

To visualize the effects of live and pasteurized A. muciniphila supplementation on microbial composition and diversity, 16S rRNA sequencing of fecal samples was performed after 15 days of gavage treatment (see Supplementary Fig. S6). At the genus level, both alpha and beta diversity showed no distinct dissimilarities were present between groups.

Discussion

Research from both our lab and others have found diminished levels of Akkermansia are associated with intestinal inflammation in both animal models and patients with IBD^27,45. Though extensive work has illustrated this microbe’s amelioratory properties in the context of metabolic syndrome^28,29,46,47, the role of A. muciniphila in the context of intestinal inflammation is still an emerging area of investigation. Here, we explored the role of A. muciniphila in intestinal inflammation and its potentially protective impact within this state by utilizing three distinct models of colitis with altered mucosal environments.

To examine the impact of A. muciniphila supplementation on intestinal inflammation, both a prophylactic and a recovery setup of the DSS colitis model were utilized. In the DSS recovery model, SPF mice were administered either live or pasteurized A. muciniphila starting day 4 of DSS administration. This treatment continued past the cessation point of DSS on day 5 and continued until sacrifice. Here, a distinct separation of DAI, particularly with regard to stool consistency and fecal blood, was apparent, with pasteurized AKK-receiving mice showing the most robust effects. This amelioration in colitis severity was mirrored in markers of inflammation, including significantly decreased levels of IL-6, TNFα, and macroscopic score. Though no significant difference in IL-10 levels were detected, comparison of the pro- to anti-inflammatory ratio²⁵ of TNFα to IL-10 revealed that in mice receiving pasteurized samples, this ratio was significantly diminished. Of note, either directly through microbial components, or indirectly through cross-feeding (i.e. via increased glycan availability and short chain fatty acid [SCFA] production)^16,21,48, A. muciniphila has been shown to influence resident microbes, a phenomenon which may contribute, in part, to its ameliorative effects in the context of colitis⁴⁹. In the recovery set up, DeSeq2 analysis (Supplementary Fig. S3) revealed that both live- and pasteurized-receiving groups, when compared with broth-receiving mice, displayed distinct blooms of Lactobacillus, a resident microbe known to have anti-colitic properties in its own right^50–53. These microbial changes emphasize A. muciniphila’s effects in vivo are not limited to interaction with the host, but manipulation of, and interactions with, other resident microbes may also play a role in its ameliorative effects. Taken as a whole, this data suggests that it may not simply be the levels of individual cytokines but the overall inflammatory milieu that influences colitis severity, and that A. muciniphila has the ability to influence the global immunological environment within the colon.

With that being said, in our work, in a prophylactic DSS experimental set-up, neither live nor pasteurized A. muciniphila-treated SPF mice showed signs of improved colitis either histologically or via cytokine analysis. In a prophylactic study set up undertaken by Bian et al.³⁰, the authors found opposing results with A. muciniphila treatment over the course of 14 days significantly ameliorated the effects of DSS-induced colitis with improved DAI, decreased proinflammatory cytokines and enhanced IL-10 in colonic tissue. It should be noted, however, that in the study conducted by Bian et al., unlike the one reported here, A. muciniphila gavage was carried out 7-days prior to DSS administration as well as during the 5-day DSS administration period, mice were taken off DSS two days prior to sacrifice, and a dose of DSS (2%) was used; these are significant procedural changes which may account for the differing results. These findings, in conjunction with our own DSS experiments, however, suggest timing of A. muciniphila administration and the inflammatory context of the gut environment are key factors in this microbe’s ability to modulate the immune system. In our work, 16S rRNA sequencing also revealed Akkermansia was unable to successfully gain a particularly robust foothold in the microbiota in this prophylactic set up. Though samples were not collected a this timepoint, it would be interesting to see if Akkermansia was able to establish and/modulate the microbiota prior to DSS administration since direct effects of DSS on Akkermansia viability, to the best of our knowledge, have not been studied.

It should also be noted that although the experiments herein were conducted using the ATCC BAA-835 strain of A. muciniphila, different strains of A. muciniphila can have significant genomic, metabolic and functional diversity as well as diverse impacts on host health, microbe-microbe interactions, and gut ecology^54–56. Indeed, based on their functional capacities, strain dependent effects of several human isolates have been reported in the context of DSS colitis⁵⁷. Similarly, a head-to-head comparison of strain 139 and ATCC-835 revealed ATCC-835 had a more robust anti-inflammatory impact in chronic DSS colitis, though both conferred beneficial effects³¹. Though this area is still underexplored, promising data in recent years has emphasized that beneficial effects in the context of colitis are observed across several Akkermansia strains^30,31,49,58 and species⁵⁹.

As mentioned above, pasteurized AKK treated mice showed the greatest ameliorative effects in the recovery model. Interestingly, a plethora of studies have shown A. muciniphila’s beneficial effects have been recapitulated, or even enhanced, with the use of pasteurized bacteria or isolated bacterial components^{15,17,22–25,47}. A. muciniphila and, indeed, several other next-generation probiotics are oxygen sensitive^17,60; therefore, finding ways to preserve this bacteria’s beneficial immunomodulatory properties while maintaining stability “off the bench” is crucial⁶¹. Pasteurization here presents a solution by adding much sought after stability, greater ease of administration, and potentially exposing the host to a wider array of intracellular components that may not be accessible in high concentrations with live bacteria, including, in the case of A. muciniphila, the outer membrane protein, Amuc_1100^24,62–65. Pasteurization may also curtail adverse effects associated with a particular microbe while still delivering the benefits⁶⁵. Our data not only provides intriguing evidence that A. muciniphila can enhance recovery from established inflammation but also suggests that these effects are promoted most effectively, in the context of recovery, by more stable pasteurized samples. These findings, most importantly, have clinical significance and may in the future aid the development of novel treatments for patients with established IBD and those patients experiencing an exacerbation of intestinal inflammation.

To expand upon our findings in the DSS model, we tested the capability of both live and pasteurized A. muciniphila to ameliorate infection severity and intestinal inflammation in the highly mucin-dependent T. muris-induced model of colitis. Mice supplemented with live A. muciniphila exhibited an environment favourable to T. muris expulsion; decreased worm burden, as well as minor up-regulation of IL-4 and IL-10, and upregulated Muc5ac were present in these mice compared to those mice receiving only broth. PAS-stained colonic tissue, however, displayed only slight signs of goblet cell hyperplasia, with regards to “full” appearance, in live A. muciniphila-receiving groups compared to both broth-receiving and pasteurized A. muciniphila-receiving groups. In addition, Muc2 gene expression did not prove significantly different between the groups in this T. muris work or the aforementioned DSS models, suggesting that a Muc2-independent mechanism may be contributing to A. muciniphila’s amelioratory effects in both conditions.

Due to the unexpected findings across DSS and T. muris experiments regarding A. muciniphila’s influence on Muc2 and PAS+ goblet cell number, we set out to investigate if the amelioratory changes observed were independent of A. muciniphila’s known influence on mucin production and turnover by utilizing a Muc2^−/− model of spontaneous colitis. Proper functioning and maintenance of the GI mucus layer is crucial in maintaining intestinal barrier function and integrity; disruption of this layer can allow the penetration of bacteria across the mucosal barrier, a major factor in the inflammation associated with colitis^1,2. Indeed, alterations in mucin quality, quantity, and mucus layer thickness are associated with CD and UC, where both goblet cells and the structural mucin, MUC2, are depleted^66–68. As mentioned above, alterations in the mucus layer are also heavily intertwined with outcomes in both the T. muris^40–42 and DSS models of colitis⁶⁹. As the name implies, A. muciniphila is intimately connected with the mucus layer of the gut. In Muc2^−/− mice, we determined that supplementation with both live and pasteurized A. muciniphila ameliorated markers of intestinal inflammation, in particular, significantly decreased levels of the pro-inflammatory cytokines, IL-6 and TNFα, in established spontaneous colitis. Though statistical significance was not reached, it should also be noted that supplementation with live AKK also conferred a marked decrease in the TNFα/IL-10 ratio, suggesting a shift to a more anti-inflammatory colonic environment in the presence of this microbe.

Of note, supplementation of A. muciniphila elicited a particular immunomodulatory pattern fairly consistent across several experiments; IL-1β was largely unaffected, whereas IL-6 and TNFα were diminished. Though these proinflammatory cytokines all play a key role in innate response and colitis progression, both IL-6 and TNFα are highly involved in modulating T cell proliferation and balancing out effector functions⁷⁰, suggesting that, in conjunction with changes in IL-10, A. muciniphila helps to shift the immunological balance in the colon, aiding in the amelioration of intestinal inflammation. Further, A. muciniphila’s protective effects do not seem to be limited by the inciting factors of inflammation (i.e. more Th1 driven in the DSS model or more Th2 dominant with respect to the T. muris model), again supporting the idea that this microbe’s mode of action in colitogenic environments is one of modulating and balancing the inflammatory milieu of the colon. In support of these findings, recent work has helped elucidate A. muciniphila’s immunomodulatory properties; Wang et al. found that administration of A. muciniphila could modulate the immunological environment in cases of colitis by reducing the infiltration of macrophages and CD8 + cytotoxic T cells to the colon³³. Several studies have also suggested that A. muciniphila has a distinct ability to promote the activity of IL-10 producing T_regs^18,26,31, potentially through the production of SCFAs^31,71,72, and/or its known interface with TLRs^25,26,73. Future research to explore and parse out the effect of IL-10 and the influence of T regulatory cells within mucin-altered environments is highly warranted.

Our investigation across these varied models of colitis supports the idea that, despite an altered mucus environment, A. muciniphila is still able to invoke anti-colitic effects and, most interestingly, is able to do so in the absence of the major structural mucin, Muc2. It should be mentioned that no overarching pattern of mucus-effecting properties prompted by A. muciniphila was detected between experiments and mixed findings with regards to Muc2 and Muc5ac gene expression suggest that A. muciniphila’s ability to influence mucin expression is highly contextual. Intriguingly, in Muc2^−/− mice, notable differences in both PAS+ stained goblet cell number and “fullness” were uncovered in the live A. muciniphila receiving group compared with the broth-receiving group. However, these same colonic tissues, when stained with Alcian blue, did not have any discernible acidic mucin-producing goblet cells. When taken as a whole, these findings suggest that live A. muciniphila, in particular, may alter the type and/or functional nature of the mucins present in the colon. A shift to more neutral glycosylation patterns may be indicative of alterations in mucin processing and glycosylation pathways in colonic goblet cells, possibly as a compensatory mechanism in Muc2-deficient environments^74–76. Notably, altered glycosylation patterns are also found in IBD, with increased terminal sialylation and fucosylation reported⁷⁷. Though beyond the scope of this work, future research exploring the impact of A. mucinphila on goblet cell biology and function, particularly with regard to those pathways involved in mucin processing and glycosylation, as well as potential compensatory mechanisms in mucin-depleted environments, is of great interest.

Collectively, these findings suggest that A. muciniphila can confer anti-colitic benefits via a Muc2-independent mechanism. These findings are particularly promising given the altered mucus environment and depleted levels of MUC2 in patients with IBD and suggest that, in the future, supplementation with this microbe may provide an additional tool for ameliorating intestinal inflammation. In addition, this work indicates that A. muciniphila supplementation may be most effective in lending an anti-inflammatory boost in established colitis. Taken together, these findings suggest a promising future for Akkermansia muciniphila as a next-generation probiotic that may help ameliorate the severity of inflammation and symptomatology in inflammatory bowel disease.

Methods

Animals

Mice were housed at McMaster University Central Animal Facility (CAF) and were kept in sterilized, filter-topped cages with autoclaved food (Teklad Irradiated Global 18% Protein Rodent Diet, 2918) and autoclaved-sterilized drinking water. All experiments were performed in accordance with the guidelines set forth by the Canadian Council on Animal Care (CCAC) and the McMaster University Care Committee, and all experiments were approved by the Animal Research Ethics Board (AREB) at McMaster University. A mix of male and female mice were used. C57BL/6 N mice (6–8 weeks old) were purchased from Taconic Biosciences (Germantown, NY) and Charles River Laboratories (Wilmington, MA). Muc2^−/− (Muc2^tm1Avel) mice (~ 12–14 weeks old) were acquired from Dr. Kris Chadee (University of Calgary)⁷⁸. These mice spontaneously develop colitis characterized by mucosal thickening and ulcerations as early as five weeks of age due to loss of epithelial barrier function⁴⁴. All mice were euthanized by manual cervical dislocation.

Bacterial culture

Akkermansia muciniphila ATCC BAA-835 was obtained from Dr. Michael Surette’s lab at McMaster University. Bacteria were prepared at 37 °C under anaerobic conditions (5% CO₂, 5% H₂ 90% N₂) in brain heart infusion (BHI) broth. Pasteurized bacteria were prepared by heating bacterial samples at 70 °C for 30 min^23,64,65. In DSS (described below) and Muc2^−/− experiments, mice were gavaged with 200 µl of 1.9 × 10⁷ CFU/ml or control broth over the course of 15 days. In T. muris experiments, resistant C57BL/6 mice were given A. muciniphila for a total of 15 days, beginning one day before infection. BHI broth was used as a control.

Chemically-induced model of intestinal inflammation

DSS (molecular mass 40 kDa; CAS NO. 9011-18-1, BOC Sciences, Shirley, NY, USA) was prepared in autoclaved drinking water to a final concentration between 1.5 and 3.5% (w/v) depending on the experimental context. Mice had ad libitum access to DSS for 5 days which replaced normal drinking water. During the administration of DSS, a disease activity index (DAI) score was used to assess the progression and severity of colitis as per our previously established methods⁷⁹. With a maximum score of 12, this scoring system is broken down into the following criteria: weight loss: 0 = no loss, 1 = 1–5%, 2 = 5–10%, 3 = 10–20%, 4 = 20%+; stool: 0 = normal, 2 = loose stool, 4 = diarrhea; and bleeding: 0 = no blood, 1 = no visual blood and Hemoccult positive (Hemoccult II; Beckman Coulter, Fullerton, CA), 2 = some blood in stool and Hemoccult positive, 3 = bloody stool and Hemoccult positive, and 4 = gross blood (blood around anus). DAI was performed by a blinded observer.

Enteric parasite infection-induced model of intestinal inflammation

T. muris is a non-invasive intestinal nematode that does not penetrate beyond the mucosal layer and provides an infection-induced model of colitis⁸⁰. In mice, within two to three weeks of infection with T. muris, the intestinal mucosa becomes thickened, crypt hyperplasia develops, and inflammatory cell infiltration is present⁸¹. Resistant mouse strains (BALB/c, C57BL/6) completely clear infection by day 35 post-infection (p.i.) with activation of Th2-type response (IL-4, IL-13)^42,81. In contrast, susceptible strains (AKR, SCID) harbour a chronic infection with activation of Th1 type response⁸¹. The ability of T. muris-infected mice to successfully clear infection is heavily reliant on changes in the mucus environment, particularly de novo expression of Muc5ac^38,39,80.

Mice were infected with one dose of approximately 300 T. muris eggs by oral gavage. The time point of sacrifice was set at 14 p.i. to assess immune alterations and cecal worm burden. As T. muris largely resides in the cecum and proximal colon, worm burden was assessed by counting the number of worms present in cecal content and tissue under a stereomicroscope.

Tissue preparation

Upon sacrifice, colonic tissue, cecal and fecal content were collected and flash-frozen in liquid nitrogen and then stored at -80℃. At the time of sacrifice, colitis severity was assessed by macroscopic scoring based on colonic and rectal bleeding, diarrhea/stool consistency, and rectal prolapse, all represented by scores ranging from 0 to 3, with 3 representing the most severe case of each condition. Each mouse was given an individualized macroscopic score based on the summation of the aforementioned criteria.

Histology

Paraffin-embedded colon sections were stained with hematoxylin and eosin (H&E) stain to assess histological damage and loss of structure, as well as periodic acid-Schiff (PAS) (as an indicator of neutral mucin-containing goblet cells) and Alcian blue (as an indicator of acidic mucin-containing goblet cells). PAS-stained and Alcian blue-stained goblet cells were quantified per 10 villus-crypt units.

H&E-stained tissue sections were analyzed by a blind observer for alterations in histology, particularly goblet cell depletion, inflammatory cell infiltration, loss of crypt architecture, muscle thickening, and the presence of crypt abscesses in mucosal tissue⁷⁹.

Enzyme-linked immunosorbent assays (ELISA)

Colonic tissue sections were homogenized in a lysis buffer (1 M Tris-HCl, 5 M NaCl, Triton X-100, dH₂O) containing protease inhibitor cocktail (PIC) (Cat. # P8340; Sigma-Aldrich, Oakville, Canada). Tissues were then centrifuged at 3300 x g for 5 min, and the resulting supernatants were stored at − 80 °C. Total protein levels of homogenates were quantified using the DC Protein Assay Kit (Cat #. 5000111; Bio-Rad Laboratories) as per the manufacturer’s protocol. Cytokine levels (IL-1β, Cat. # MLB00C; IL-6, Cat. # M6000B; TNFα, Cat#MTA00B, IL-13, Cat. # M1300CB; IL-4, Cat. # M4000B) were determined according to the manufacturer’s instructions (Quantikine Murine; R&D Systems, Minneapolis, MN, USA). The anti-inflammatory cytokine, IL-10, was measured using a commercially available high-sensitivity ELISA kit (Cat. # EKN49384; Biomatik, Kitchener, Canada).

Quantitative reverse transcriptase polymerase chain reaction (RT-qPCR)

In colonic tissue, RNA extraction was performed using TRIzol^® reagent (Ref. # 15596018; ThermoFisher Scientific/Ambion by Life Technologies, Carlsbad, CA, USA). The forward and reverse primer sequences can be found in Supplementary Table S1. RT-qPCR was executed using the CFX96 real-time PCR system (Bio-Rad). The resultant data were analyzed using the 2^−ΔΔCT method. Results are expressed relative to the housekeeping gene coding for 18S ribosomal RNA (rRNA).

Microbial profiling and analysis

To analyse the composition of the microbiota and changes in diversity with A. muciniphila supplementation, 16S rRNA sequencing of fecal samples was performed via amplification of the V3 regions. Amplification products were sequenced via Illumina MiSeq Illumina as previously described⁴⁵, processed via the DADA2 pipeline and compared against the SILVA taxonomy database. Microbial data were analysed using the online tool MicrobiomeAnalyst developed by the Xia Lab at McGill University.

The impact A. muciniphila supplementation exhibited on the bacterial community was visualized by Bray-Curtis beta diversity, and statistically significant differences were examined using permutational multivariate analysis of variance (PERMANOVA). To account for both evenness and richness, alpha diversity was analysed utilizing the Shannon Index.

Student’s t-test and one-way analysis of variance (ANOVA) were used, where appropriate, for univariate statistical comparison. P values were adjusted using false discovery rate (FDR). Corrected P values of < 0.05 were considered statistically significant.

Statistical analysis

Statistical analysis was completed using GraphPad Prism version 9.0.0 for Windows (GraphPad Software, San Diego, California, USA). All results are expressed as mean ± standard error of the mean (SEM). Comparison between groups was performed using one-way analysis of variance (ANOVA) with Tukey’s multiple comparison test, where applicable. P values of < 0.05 were reported as statistically significant.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

J.A.G. and W.I.K. designed the experiments. J.A.G., S.H., H.W., C.C., H.D. and M.G.S. performed the experiments and analyzed the data. J.A.G. wrote the manuscript and designed and created the figures. L.M. and M.G.S. provided invaluable feedback for the study design and analysis. W.I.K. supervised the project.

Funding

This work was supported by grants from Canadian Institutes of Health Research (CIHR; PJT 156262) and Natural Sciences and Engineering Research Council of Canada (NSERC; RGPIN-2019-06739) to WIK, and by the Canadian Graduate Scholarship—Master’s program (CGS-M) to J.A.G.

Data availability

The datasets generated and analyzed for the current research are available from the corresponding author upon reasonable request. 16 S rRNA datasets generated and analysed during the current study are available in the Sequence Read Archive under accession number PRJNA1447570 (Reviewer link: [https://dataview.ncbi.nlm.nih.gov/object/PRJNA1447570?reviewer=mqscsjf5ddppvsbrgo1vuhp00p])

Declarations

Competing interests

The authors declare no competing interests.